‘The effects of using English in online
advertising’
- an eye tracking study into the effects of the use of
English in Dutch online product advertisement banners
Master’s thesis: International Business Communication
Sanne Handgraaf, S4070860 sannehandgraaf@student.ru.nl
Supervisor: Dr. U. Nederstigt Assessor: Dr. B. Hilberink
Master program of International Business Communication
Faculty of Arts
Radboud University Nijmegen
2
Abstract
As online advertisers are competing for the limited attention of their audience, online advertisements need to attract as much attention as possible and need to create positive attitudes toward the advertised product and brand. It has been suggested that the use of foreign language can improve the overall effectiveness of advertisements. As the English language occurs frequently in non-English advertising, this study investigated the effect of English in Dutch online advertising with regard to several aspects: the effectiveness of drawing attention, the attitude toward product, brand and advertisement, the purchase intention, the recall of product, brand name and advertising slogan and the appreciation for the language used. An experimental eye tracking study with a mixed design was carried out among 60 participants, who each saw six webpages that contained one banner advertisement and information on a job vacancy. The banner advertisement contained either Dutch or English slogans. The participants were instructed to look for specific information about the job vacancies. This goal-directed task was given in order to produce looking behaviour that would closely match real-life looking behaviour of browsing Internet users. An eye tracker registered the looking behaviour of participants and a questionnaire was used to measure attitudes, recall and appreciation. The results imply that the effectiveness of English advertisements does not differ from that of Dutch advertisements. Even though the
participants reported to appreciate English advertisements more than Dutch advertisements, an actual effect was lacking. Strikingly, almost no attention was paid to the banner
advertisements at all. The findings of this study can be practically relevant for online advertisers, as they imply that the effect of English does not differ from the effect of Dutch with regard to attracting attention and that the use of English does not cause more negative or positive attitudes. Advertisers can consider these findings when using English in non-English advertising in order to standardize their international marketing strategies and to lower translation costs.
Further research could focus on the discrepancy between the perceived preference of English advertising and the lack of actual effects on attitude. The questions raised in this study about the effectiveness of online banner advertisements could be answered in further research. Finally, additional insights into the field could be yielded by investigating the effect of different foreign languages on online advertising.
3
Introduction
The abundance of information on the Web and the competition for the limited attention of Internet users have led to the so called “war for eye balls”, especially among advertisers (Schiessl, Duda, Thölke & Fischer, 2003). Online advertisements have an
immense potential audience, but in order to reach this audience, the online ads have to attract as much attention as possible.
It has often been suggested that the use of foreign language can improve the effectiveness of advertising (Domzal, Hunt, & Kernan, 1995; Gerritsen, Korzilius, Van Meurs, & Gijsbers, 2000; Hilgendorf, 1996; Hornikx & Starren, 2006; Piller, 2001; 2003) but there is only little empirical evidence for the effects of foreign language in advertising. An adequate way of investigating attention toward advertisements is eye tracking methodology (e.g. Rayner, Miller, & Rotello, 2008; Toebes, 2013). Eye tracking research measuring the effects of foreign language use in online advertising is lacking entirely. This experimental eye tracking study investigated the effect of English in Dutch online advertising, focusing on the following aspects: the effectiveness of drawing attention, the attitude toward product, brand and advertisement, the purchase intention, the recall of product, brand name and advertising slogan and the appreciation of the language used.
Literature Review
The use of foreign languages in advertisements
The use of foreign languages has been observed in many different kinds of advertising messages, in many different countries (Gerritsen et al., 2007; Griffin, 1997; Hilgendorf, 1996; Pétery, 2011; Piller, 2001; Taavitsainen & Pahta, 2003; 2001; Ustinova, 2006). Especially English is used as a non-native language in advertisements and has become the most
frequently used foreign language in advertisements in non-English speaking countries (Piller, 2003). In a corpus of 2384 advertisements from several European countries, Gerritsen et al. (2007) found that two-third of the advertisements contained English words. The Netherlands, also part of this study, showed a comparable high number.
4 Several studies have attempted to explain this frequent occurrence of foreign language in advertising. The suggested reasons of the use of foreign language in advertising can be divided into symbolic and non-symbolic purposes (Van Meurs, 2010). The symbolic purpose of a foreign language is to evoke certain connotations that the reader has with the language. Piller (2003) argued that advertisements that include elements in a specific foreign language evoke associations that the reader has with the people and the culture from the country where this foreign language is spoken. The use of French would evoke associations with love and Italian expressions would evoke associations with a good life and quality food (Piller, 2001). Several studies have suggested that these connotations play a role in advertising with foreign language elements (Haarmann, 1989; Kelly-Holmes, 2000; Piller 2001; 2003; Warden, Lai, & Lu, 2002). According to Hornikx and Hof (2008) advertisements can benefit from this if the evoked associations and images are actually relevant to the product that is advertised.
However, English is rarely used to create country-specific associations. The strength of English seems to be its status as a world language and the competence this entails (Van Meurs, 2010). According to Piller (2001), the values that are associated with English are globalization, modernity and progress. Additionally, Hornikx and Starren (2006) implied that the use of English gives a company or brand an international and contemporary orientation. It has been suggested by De Raaij (1997) that the use of English in Dutch advertisements makes advertising texts appear more intelligent and interesting. It is possible that readers attribute the associations they have with English, for instance those of globalization, modernity and
progress, to the entire advertisement or even to the brand behind the advertisement. The idea that the use of English could affect the attitude towards a brand is particularly interesting, as it was shown by Brown and Stayman (1992) that attitude toward a brand significantly impacts behavioral intention. Brown and Stayman claimed that there is a positive relationship between attitude towards brand and intention to buy.
Non-symbolic purposes of using foreign language in advertising are those that have nothing to do with the connotations that specific foreign languages evoke, but that have a practical background. A highly practical motivation for internationally operating companies to use English is the standardization of their marketing strategy (De Mooij, 1991). Using the same language throughout a multinational or global advertising strategy can lower costs and ensure consistency (Agrawal, 1995; Taylor & Okazaki, 2006; Warden et al., 2002). It makes translations of advertising phrases for every country that a company operates in unnecessary and it gives the company an opportunity to create a global image that is the same across these countries. Furthermore, in some cases no equivalent translation can be found for an
5 advertising phrase in the mother tongue (L1) of another country. De Raaij (1997) observed that the use of English elements in Dutch advertisements can add value in the case that the Dutch translation would be too long, as the English language contains many monosyllabic words. In the case of some products, advertisers want to address diverse target groups and cannot decide whether to approach the recipient with the formal “u” or the informal “jij”. In all of these cases, the use of English words or expressions could be a solution.
Most importantly, it has been suggested by many studies that foreign language attracts more attention of the reader than the use of native language does (e.g. Domzal et al., 1995; Gerritsen et al., 2000; Hilgendorf, 1996; Piller, 2001; Takashi, 1990). Domzal et al. (1995) claimed that the use of another language than the L1 of a country makes advertisements more distinctive as people do not expect to come across a foreign language. This idea was already supported in 1979 by Bettman, who claimed that distinctive information draws more attention than ordinary information. According to this theory, advertisements with foreign expressions should be noticed more, but should also be processed more deeply which leads to a higher level of recall. Most readers would need more time to read a foreign expression, as it is more complex to understand than an expression in the L1. Following this, advertisements that include foreign expressions possibly attract more attention and are processed deeper. Subsequently, they are remembered better than the advertisements without foreign
expressions. The distinctive appearance of foreign language should additionally cause the readers to look at the advertisements with foreign language first.
Apart from the benefits that the use of foreign language in advertising has, there are also possible obstacles when using a foreign language strategy. Gerritsen et al. (2000) found that the use of English in Dutch television commercials could lead to miscomprehension and in frequent cases to a negative attitude towards the product or brand advertised. Hornikx and Starren (2006) additionally found that easy advertisement slogans were more appreciated than the difficult slogans. These findings show that, if not carefully adapted to the audience, the use of foreign language could also achieve the exact opposite effect of what it was primary meant for, namely a negative attitude toward the product and brand advertised. Another obstacle that could occur is that the processing of a second language (L2) puts more pressure on a recipient than the processing of an L1 (Luna & Perachhio, 2002). The foreign expression should not demand too much effort from the recipient, since this will lead to frustration or worse (Domzal et al., 1995).
The literature that has been discussed so far shows the prevalence of research into print advertising. However, a change is taking place and as the use of print advertising is
6 decreasing, online advertising is ever growing (Barnard, 2015). Print advertising and online advertising are comparable with regard to some aspects: text and graphic display of the Web are similar to that of print and the usage of the Internet has the same “paced,
self-selective” nature (Ha, 2008). Therefore, the previously discussed literature offered valuable insights for the present study. However, it is important to take into account the distinctive features of online advertising as well. For that reason, the following paragraph will outline the distinctiveness of online advertising and the role that foreign language could potentially play here.
Online advertising and the use of English
In the past decadesthe Internet has become the fastest growing advertising medium, with a still growing number of users around the globe (Ha, 2008). According to forecasters, the Internet will be the biggest medium in several global ad markets by 2017, taking over the current biggest advertising medium of television by 2020. Internet advertising will continue to increase rapidly, while the use of print advertising media such as newspapers and magazines will continue to decrease (Barnard, 2015). Since the 90s there has been specific attention for research into this new genre of advertising that the Internet has brought us (e.g. Bayles, 2002; Berthon, Pitt, & Watson, 1996; Burke, Hornof, Nilsen, & Gorman, 2005; Danaher &
Mullarkey, 2003). However, there is still a large gap in the research that acknowledges the distinctive properties of online advertising as opposed to the traditional form of print advertising (Ha, 2008). As online advertising is still growing, the demand for effective advertising and media planning strategies for online advertising in particular is increasing (Danaher & Mullarkey, 2003).
Berthon et al. (1996) were the first to publish an article about the Internet as an
advertising medium and praised the Internet for being a two-way communication medium that is not time-bound and can take place regardless of the distance between the users. They called the Internet a unique medium because participation is relatively easy and cheap and because the Internet is international by definition. According to Coyle and Thorson (2001) the Web distinguishes itself as an advertising medium by its vividness and interactivity that leads to differences in how advertisements are processed.
A great distinctive feature of the Web is that it offers the opportunity to integrate more stimuli, such as animations and interactivity, than would be possible for traditional genres of advertising. Dahlen, Murray and Nordenstam (2004) suggested that with these tools, online advertisements could attract significantly more attention. Several studies have claimed that
7 these stimuli (e.g. the interactivity and animations) increase the attention for the ads and that these stimuli lead to higher click-through rates (Kimelfeld & Watt, 2001; Sundar & Kim, 2001).
Within research into attention drawing aspects of online advertisements, it should be kept in mind that stimuli, such as animations and interactivity, indeed play a leading role. Nevertheless, additional factors, such as the use of foreign language, could still contribute to the effectiveness of attention drawing and are therefore well worth investigating too.
The role of language on the Internet seems greatly determined by the international reach that the Internet has (Berthon et al., 1996), considering that public webpages are accessible from any location in the world. For advertising, this has several implications related to the language used online (e.g. Warden et al., 2002; Ha, 2008). Warden et al. (2002) mention the technical problems for firms to include multiple languages on a single website and report a prevalence of English on websites that target Asian consumers. Ha (2008) signals that standardization and localization issues play a large role in determining online language strategies for marketers.
Figure 1. Two Dutch online advertisements on an American website that are personalised with the use of demographical data
The local orientation of firms can be conveniently achieved through Online Behavioral Advertising (OBA). OBA is also called “Internet-based advertising” and is a tool for
advertisers to organize their online audience so they can target readers with relevant
8 by using their assumed demographics, interests, location or online behavior (The Digital Advertising Alliance of Canada, 2015). OBA is operationalized through the use of cookies that collect information from the websites Internet users visit, computer settings and entries in search engines. This enhances possibilities for advertisers to personalize advertisements, for instance by showing local product offers. An example of this can be seen in Figure 1.
An implication of OBA is that online advertisements are generally targeted at Internet users in the language that is connected to their IP address or computer settings. This language is most likely to be the L1 of the Internet user and a Dutch audience would therefore generally receive Dutch advertisements. However, there are exceptions to this condition, for example in the case that a non-Dutch website does not implement OBA and shows standardized advertisements in the language of the website. In such cases, a Dutch audience would receive English advertisements when viewing these particular websites. The experiment of the present study only used Dutch websites to ensure that the English element used in the advertisements remained distinctive.
The discussion on language use in online advertising raises an important question for the present study: in which language does the audience prefer to be targeted by advertisers? The literature that has been discussed so far leads to two different perspectives: the Internet can be considered an international medium without ‘boarders’, where a majority language such as English might be the best language to approach Internet users. However, at the same time the Internet allows for personalization of advertisements, which could lead to a
preference for a personalized, native language approach. To answer this question, this study investigated both the different effects of the use of English and Dutch in advertising on attitude toward product, brand and advertisement, and also the language that Dutch Internet users prefer to be targeted by, by advertisers.
In order to investigate online advertising, the format of online advertisements needs careful consideration, given that multiple variations can cause different attitudes among Internet users (Burns & Lutz, 2006). Of the six main categories of formats that Burn and Lutz studied, consisting of banner advertisements, floating ads, large rectangles, interstitials, skyscrapers and pop-ups, banner advertisements were shown to be the most informative and caused a more positive attitude from Internet users. Banner ads are horizontal, rectangular shaped advertisements found at the top of webpages. In addition, banner ads scored highest in all the behavioural measures, such as the click through percentage, the percentage of later visits and the click through frequency, even though Burke et al. (2005) found that viewers of a website consciously try to avoid banners. Finally, Burns and Lutz found that banners on top of
9 a webpage were remembered better than banners that are placed inside the contents of a webpage, even though viewers looked less frequently at the banners on top of the page. Supported by these findings, this study used banner ads for the experimental design.
An additional factor that was considered in this study was that some Internet users make use of advertisement blocking software. This software can block the majority of online advertisements, although companies such as PageFair and Adobe are already investigating ways to overcome this obstacle for advertisers. In 2014, these companies published a report on the growth of adblocking software users. It was announced that 4.9% of all Internet users employed adblocking in 2014 and that this percentage is increasing rapidly. The reasons for Internet users to block advertisements are the lack of desire to view any advertising, privacy concerns and annoyance with intrusive advertisement formats. So far, the effect of adblocking on the attention for online advertisements is unknown. Therefore, the effect of the use of adblocking software on the attention of Internet users was investigated in this study as well.
In order to measure the attention of Internet users, eye tracking research was deployed. In the following paragraph, the choice for this particular methodology will be outlined.
Eye track to measure attention
Eye tracking methods are used to measure where a person is looking at any given moment and the sequence on which the eyes are moving across the screen (Poole & Ball, 2006). The method has been used before to gain insight into the effectiveness of online contexts, as it offers the opportunity to evaluate online designs by comparing patterns of fixation with areas of interest on websites (Sutcliffe & Namoune, 2007; Wedel & Pieters, 2007).
So far, eye tracking research has led to findings about several visual attention aspects. For example, Outing and Ruel (2004) found empirical evidence for the claim that Internet users’ point of entry of a website is the upper left corner, usually the headline, by tracking their eyes for initial viewing behavior. Sutcliffe and Namoune (2007) investigated the effect of website structure and layout on the user attention and user interest. Layout was shown to play an important role in the ability of a website to draw attention effectively: websites with an overall open layout, as opposed to a traditional columnar block layout, were rated more attractive and websites with column layout were preferred. These examples show that eye tracking is a suitable method of studying visual attention on the screen, since “attentional processes do not depend solely on conscious control” (p. 2) and therefore cannot be only measured through self-reported attention (Schiessel, 2003).
10 There are different methods of measuring eye movement, but this study concentrated on fixations. Fixations take place when the viewer focuses on one point for a relatively long time. Therefore, when a viewer is fixating on a specific area, it can be assumed that the viewer is processing the information. The more fixations a particular area receives, the greater the interest for this area is (Poole & Ball, 2006). Poole and Ball name several fixation metrics of which four will be used in the present study. The number of fixations per area of interest indicates whether a particular area is more noticeable or important than other areas. The duration of fixation indicates whether an object is more engaging or more difficult to
comprehend. In addition to the number of fixations per area, this study took into consideration the repeated fixations on the area of interest and the first fixation. Betmann (1979) gives a relevant example of the importance of fixation as an eye tracking metric. He states that deeper processing of information should lead to significant longer fixation duration on the area of interest. Betmann clarifies why eye track research is so valuable in the field of advertising research: he suggests that deeper processing of information should lead to significant longer fixation duration on advertisements with foreign language elements as well. The distinctive appearance of foreign language should moreover lead to more first fixations (Betmann, 1979).
An eye tracking study by Toebes (2013) investigated the effect of English and Italian language use in Dutch print advertisements on attention, attitude and purchase intentions. He found that advertisements that used English or Italian did not receive more attention than Dutch advertisements, except for the measures of number and duration of fixations: English slogans did seem to be more effective at attracting attention. The attitude and purchase intentions towards advertisements were lower when the advertisements contained Italian, compared to English and Dutch. Toebes concluded that English makes a better alternative for Dutch than Italian. The study by Toebes was an important starting point for the current study, as this study made use of a similar methodology and adopted the materials that were used by Toebes. The current study also used eye tracking to investigate the effectiveness of the use of English in Dutch online advertisements with regard to attention drawing.
For Internet advertising research, the instruction that participants receive when viewing a website in an eye tracking research matters greatly. The goal of the viewer appears to have large implications for looking behavior, since it influences where viewers look (Rayner et al., 2008). Wedel and Pieters (2007) found that the instructions reduced the amount of attention for the advertisements used in their study. Research by Danaher and Mullarkey (2003) showed that participants were more involved when given a goal-directed task, which was likely to reduce the impact of peripheral materials such as advertising. When the participants
11 were instructed to explore a webpage there was more attention for banner advertisements. In this study, participants were given a goal-directed task. This created a realistic and natural situation in which participants were browsing the Internet for the information that they were looking for, as they would do when surfing the Internet under usual circumstances.
The literature discussed has led to the following research questions:
RQ1: Is the use of English in online advertising more effective at attracting attention than
the native language?
RQ2: Is there a difference between the attitudes toward the advertised product, the
advertised brand and the advertisement between online advertisements that contain English and online advertisements that only contain the native language of the audience?
RQ3: Does the recall of product and brand name differ between online advertising that
contains English and online advertising that contains only native language elements?
RQ4: Does appreciation differ between online advertising that contains English and online
12
Method
An experiment was conducted to investigate the effect of English in Dutch online advertisements on attention, attitude and recall.
Design
For the experiment, a mixed design (within-subject and between-subject) was used. Four groups of participants were shown six webpages, each containing a job advertisement and a banner ad (see Appendix A). The within-subject factor was the language of the slogan of the banner ad. This slogan was Dutch for three ads and English for the other three ads. Two of the four groups saw the same ads as the other two groups, but with the opposite language integrated in the slogans. Additionally, the sequence of the ads was reversed in each set of groups. The sequence was not considered in the analyses1.
The colors and position of the components of the banner ads were systematically varied to minimize the influence of these factors.
Materials
Because this study aims to study online advertising, advertisements were integrated in webpages that offered a rich context of information. As stated in the introduction, banner ads placed on top of a webpage scored highest in behavioral measures such as click through rates, but also in recall (Burns & Lutz, 2006). Therefore, this study used banner ads that were placed on top of the webpage. The ads were fabricated especially for this experiment and highly simplified to minimize differences in context, text and visuals, as in Toebes (2013). No animations were used in the banner ads for the same reason, but also because the experiment building software would not allow for analysis of these animations.
Components of the ads were adapted from the material from Toebes (2013), as this material was carefully developed and thoroughly pretested with regard to several factors that will be outlined in the following paragraph.
In the advertisements used by Toebes, the foreign language of English only occurred in the slogan of the ad to allow for better control. The rest of the textual elements on the
13 webpage were in Dutch, in order to ensure that attention would be drawn by the language of an element and not by irrelevant aspects. The products that were featured in the ads were electronic products (televisions, laptops, tablets, MP3-player, smart phones and photo cameras) because the use of foreign language is especially popular in the advertising of this market (Alm, 2003; Gerritsen et al., 2007). The pictures that were featured in the ads were of electronic goods that were not sold in the Dutch market. This way, Dutch participants would not be likely to recognize the products. Products that are known by the participants run the risk of influencing attitudes toward the product due to preexisting experience with the product. Furthermore, non-existing brand names were created by an online non-word generator and pre-tested to check for familiarity. No non-words were recognized, but words that evoked associations referring to objects or products were taken out. For the slogans, Toebes used soft sell appeals, as these are most frequently used in Dutch advertising
(Koeman, 2007). The length of the slogans was based on existing Dutch slogans for electronic product ads, namely between 23 to 39 characters. The accuracy of the Dutch and English slogans was tested by Dutch non-native English speakers.
The slogans, the pictures of the products and the body texts with the technical
specification of the advertised products were adopted from Toebes (2013) and the final design was inspired by existing ads. These components were used in the design of the banner ads that was inspired by existing banner ads. An example can be seen in Figure 2. Each banner
contained a picture, a slogan, body text with technical specifications of the product, and a button, an element that is frequently observed in online advertising. The colors and the location of the components were systematically varied. The colors used in the banner ads (grey/black and orange/white) were inspired by the colors that were used in actual banner ads found on the Internet.
Each banner ad appeared on a webpage that was part of a job vacancy website
(Appendix A). Every webpage contained a different job advertisement that was targeted at an audience with an applied university (HBO) or academic university (WO) background in order to match the background of the participants of this research. Each job advertisement contained the logo of the organization that was offering the vacancy. The six different webpages
differed in length and showed different aspects of job advertisements, as the participants were instructed to look for specific information about the jobs that were advertised.
A questionnaire was made in fourfold, to match the four different versions of the experiment (Appendix B).
14 Figure 2. Banner ad 1 with English slogan.
Participants
A total of 61 participants took part in this study. One participant was excluded from the dataset due to unusable eye track data. Therefore, the analyses in this study were conducted for the 60 remaining participants. All the participants were native speakers of Dutch. 50 participants studied or graduated from an academic university (WO) and 10 participants studied or graduated from an applied university (HBO). Twenty-four participants were male (40%) and 36 were female (60%). The mean age of the participants was 23.72 years (SD = 5.16), ranging from 19 years old to 56 years old. The mean perceived English reading
proficiency, on a scale from one (very poor) to seven (very good), of the participants was 5.72 (SD = 1.04). The overall English proficiency, on a scale from one (very poor) to seven (very good), of the participants was 5.36 (SD = 0.94). Overall English proficiency consisted of the self-assessed levels of reading, listening, writing and speaking in English.
All 60 participants stated that they use the Internet on daily basis. Forty-two of the 60 participants (70%) reported that they make use of adblocking software on their private laptop or personal computer.
Independent-samples t-tests were conducted to see whether there was a difference between groups A (exposed to version 1 and 3 of the experiment) and B (version 2 and 4) in terms of age and perceived English proficiency.
The independent-samples t-tests revealed no significant differences for age between the two conditions (t (58) = 1.53, p = .131) and for perceived English proficiency between the conditions (t (58) = 1.17, p = .247). Gender distribution, educational background and the use of online advertising blocking options were tested per condition by square tests. The Chi-square analysis revealed no significant relation between condition and gender (χ2 (1) = 0.10, p = .752). Additionally, no significant relation between condition and educational background was found (χ2 (1) = 0.65, p = .419). Finally, a Chi-square analysis showed no significant relation between condition and the use of online advertising blocking options (χ2 (1) = 0.54, p = .464).
15 Instruments
Several variables were measured in this study, namely the four different eye tracking measures of first fixation, fixations per area of interest, duration of fixation and return of views on the area of interest, the perceived English language proficiency, attitudes toward product, brand and advertisement, purchase intention, recall and, at last, self-reported appreciation.
In order to measure attention, the EyeLink II by SR Research was used. This eye tracker consists of three miniature cameras that are secured on a comfortable padded headband that was mounted on the heads of the participants. The eye-movements of participants were recorded with the use of the software of Experiment Builder by SR
Research. Attention was measured in terms of four different factors (Poole & Ball, 2006): first fixation, fixations per area of interest, duration of fixation and return of views on area of interest. First fixation is the first element of the webpage that a participant fixated on when they were shown the entire webpage. Since participants were shown an instruction in the first nine seconds of each trial, fixations that were recorded during those first nine seconds of each trial were deleted. Fixations per area of interest is the total number of fixations on the specific areas that were relevant in this research, such as the banner ad and the job advertisement. The proportion of duration of fixation on area of interest is the relative length of fixations on each area of interest. The return of views on area of interest is the number of times participants’ fixations go back to the area of interest.
For the eventual analyses for attention, the data for the different advertisements and trials were merged.
The perceived English language proficiency was measured through self-assessment
of language skills. Participants had to indicate how good their writing, speaking, listening and reading skills in English were, ranging from 1 (very poor) to 7 (very good). The reliability of perceived English proficiency was good (α = 0.81).
Additionally, the constructs of attitude, recall and appreciation were measured. Attitude was measured through several subconstructs, namely attitude toward the advertised product, attitude toward the advertisement, attitude toward the brand and the purchase intention. All subconstructs were based on Hornikx, De Groot, Timmermans, Mariëns and Verckens (2010).
When filling out the questions regarding attitude and purchase intention, the banner ads were shown to the participant again. This time, only the banner ads were shown.
16 The following paragraphs outline the different items that were measured. After the reversed items were recoded, the scales could be interpreted as follows: a low score indicates a negative attitude with regard to the item and a high score a positive attitude.
The attitude toward advertised product was measured with seven-point Likert scales, consisting of five items (‘appealing-unappealing’, ‘beautiful-ugly’, ‘good-bad’, ‘low quality-high quality’, ‘not interesting-interesting’) of which ‘beautiful-ugly’ and ‘good-bad’ were reverse coded in the questionnaire. Cronbach’s alphas were used to measure the reliability of all constructs per version. The reliability for attitude toward advertised product for Dutch slogans was acceptable (α = .69) to excellent (α = .97) and for English slogans the reliability was acceptable (α = .71) to excellent (α = .96). See Appendix C for all Cronbach’s α’s for attitude toward product, attitude toward advertisement, and purchase intention, as a function of the product and language.
The attitude toward the advertisement was measured with seven-point Likert scales, consisting of five items (‘not interesting-interesting’, ‘unappealing-appealing’, ‘beautiful-ugly’, ‘discrete-notable’, pleasant’), of which ‘beautiful-ugly’ and ‘unpleasant-pleasant’ were reverse coded in the questionnaire.The reliability for attitude toward the advertisements with Dutch slogans was low (α = .592) to excellent (α = .93) and for
advertisements with English slogans the reliability was acceptable (α = .76) to excellent (α = .92).
The attitude toward the brand was measured with seven-point Likert scales, consisting of three items (‘unreliable-reliable’, ‘competent-incompetent’, ‘bad-good’) of which ‘bad-good’ was reverse coded. The reliability for attitude toward brand was acceptable (α = .65) to excellent (α = .99) for advertisements with Dutch slogans and acceptable (α = .75) to excellent (α = .96) for the advertisements with English slogans.
The purchase intention was measured with a seven-point Likert scale, consisting of a statement (‘I am considering purchasing the product that is featured in this advertisement’) to which participants could agree or disagree (‘highly disagree-highly agree’).
For the eventual analyses for attitude and purchase intention, the data for the different advertisements and trials were merged.
Recall was measured through a free recall memory test (Sutcliffe & Namoune, 2007).
Since this experiment was part of a larger research project, participants took part in a different
2 A Cronbach alpha value of .59 is low and would not be acceptable under normal circumstances. However, in
this case there was only a small amount of items which does allow for one low Cronbach alpha value (Briggs & Cheek, 1986).
17 eye track experiment instantly after looking at the webpages of the present study. This
ensured a reasonable amount of time between seeing the ads and filling out the free recall memory test. Participants were asked to list the products, brands and slogans that they could remember for each webpage. The company and the position that featured in the job
advertisement were given to the participants by means of a small amount of support. The responses of the participants were coded by two independent coders.
Finally, self-reported appreciation was measured by asking the participants about their general appreciation for advertisements in the Dutch language and advertisement containing English text elements. The constructs were measured with seven-point Likert scales, consisting of five items (‘not-interesting-interesting’, ‘unappealing-appealing’, ’ugly-beautiful’, ‘discrete-notable’, ‘unpleasant-pleasant’) of which two items were reverse coded (‘ugly-beautiful’, ‘unpleasant-pleasant’). The reliability of appreciation for advertisements in Dutch (α = 0.87) and the reliability of appreciation for advertisements containing English (α = 0.85) were both good.
Procedure
Participants were recruited in the area of Nijmegen and surroundings and took part in the experiment during the month of June. The experiment of the current study was part of a greater study and therefore there were several additional components included in this experiment that had no relevance to this study. Participants were systematically assigned to one of the four subgroups. First, they were seated in front of the computer screen with a resolution of 1920 by 1080. All participants were asked to sign a consent form. The
participants were informed that the aim of the research was to investigate the findability of information on a new job advertising website. The instructions appeared on the computer screen and were later summarized by the researcher. During the preparations for the experiment, all participants had the opportunity to ask questions.
As the task given to a participant is highly important for looking behavior (Rayner et al., 2008), the present study aimed to set up an experiment that closely matched a real-life situation in which Internet users browse the Internet for information. Therefore, the
participants were given a goal-directed task. Before looking at each webpage, the participants were shown a question about the contents of the job advertisement for nine seconds, for example: “What distinguishes this employer from other employers?” The participants were then allowed to look at the website and answer the question out loud. The time spent looking
18 at the webpage was determined by the participants themselves, as they could use the computer mouse to close the webpage when finished. This process was repeated for six webpages.
When the participants were ready to start, the eye tracker was attached to their heads and the cameras were adjusted into the right position. The eye tracker was then calibrated and validated before using a practice trial to make the participant comfortable and familiar with the procedure. After the practice trial, the participant had the opportunity to ask more
questions to the researcher. Before the actual experiment started, the eye tracker was validated and calibrated once more.
When the eye track part of the entire experiment finished, the participant was asked to fill out the questionnaire (Appendix B). The average total duration of this experiment was 35 minutes.
By means of an incentive, participants had a chance to win four vouchers of ten Euros and one voucher of twenty Euros.
Since an occasional poor quality of calibration and/or validation caused errors in the data, the fixation points outside of the interest areas were manually altered when all experiments were completed. Nine trials had to be excluded because a significant fraction of the points were outside of the interest areas. Subsequently, the number of trials decreased from 360 to 351 (N = 351).
Statistical treatments
When all the data was retrieved, several statistical tests were conducted. Independent t-tests and Chi-square analyses were conducted to test the differences between the conditions.
In order to analyze the eye track data, repeated measures ANOVAs were performed to compare fixations per area of interest, proportions of duration of fixation on area of interest and the return of views on the area of interest. The first fixations were analyzed using a Chi Square analysis. The effect of the use of advertisement blocking tools was analyzed using one-way ANOVAs.
The analyses for the several constructs of attitude and for recall were conducted with the use of repeated measures ANOVAs. The relation between the use of advertisement blocking software and the overall attitude toward advertisements was analyzed with a simple regression analysis.
19
Results
The purpose of this study is to investigate the effects of the use of English in Dutch online advertisements with regard to several factors: the effectiveness of drawing attitude, the attitude toward product, brand and advertisement, the recall of product, brand name and advertising slogan and the appreciation of the language used. In this results section, several analyses were conducted to investigate these differences. First, the eye tracking data was studied and subsequently the data that was retrieved from the questionnaires was studied.
Eye tracking
1. First Fixation
One of the ways attention was measured in the present study was through first fixation. The first research question concerned whether there would be a difference in terms of initial fixation on the banner ad, when the slogan of the banner ad was in Dutch or when the slogan was in English.
Firstly, an analysis was performed to find out if there was a difference in first fixation for the different webpages that participants viewed. A Chi-square analysis revealed no significant relation between webpage and first fixation (χ2 (15) = 18.68, p = .185), which indicates that all banners drew attention in a similar fashion (see Table 1).
Table 1. The total number of first fixations, specified per webpage per relevant area of interest (N = 351, due to the deletion of trials, nine trials were excluded)
Position of first fixation
Banner Vacancy Answer Other Total
Webpage 1 5 50 0 1 56 Webpage 2 5 50 0 3 58 Webpage 3 3 49 0 6 58 Webpage 4 6 53 0 0 59 Webpage 5 3 56 0 1 60 Webpage 6 3 55 1 1 60 Total 25 313 1 12 351
20 Subsequently, an analysis was conducted for the differences between first fixation for Dutch and English advertisement slogans. Also here, the Chi-square test revealed no significant relation between language of the banner ad slogan and the first fixation (χ2 (3) = 1.04, p = .791). No relationship was found between first fixation and the language of the banner ad slogan.
Tables 1 and 2 shows that the first fixation was in the contents of the webpage in a great majority of the trials. This was the area where the information could be found that the participants needed for fulfilling their task. Only in a few cases, 25 out of 351 (7.12%), participants looked first at the banner ad on top of the webpage.
Table 2. The total number of first fixations, specified per language per relevant area of interest (N = 351, due to the deletion of trials, nine trials were excluded)
Position of first fixation
Language of advertising slogan Banner Vacancy Answer Other Total
English 13 158 0 6 177
Dutch 12 155 1 6 174
Total 25 313 1 12 351
2. Fixations per area of interest
Attention was also measured in terms of the number of the fixations on the banner ads. The first research question addressed the relationship between the number of fixations on the banner ad and the language of the banner ad slogan. A repeated measures analysis for number of fixations on the banner ad with as factor Language of the advertisement slogan showed no significant main effect (F (1, 59) = 0.01, p = .981). No relationship was found between the number of fixations on the banner ad and the language of the banner ad slogan.
Table 3. The means and standard deviations for the number of fixations on the banner ad, specified per language of the banner ad slogan (N = 60)
Language advertising slogan M SD
Dutch 1.03 1.97
English 1.03 1.79
As the use of advertisement blocking software could have a possible effect on viewing behaviour of Internet users, and thus on the number of fixations on the banner ad, the
21 relationship between the use of adblocking software and the number of fixations on the
banner ad was investigated. A one-way ANOVA with as factor Use of advertisement blocking software (using/not-using) showed no significant main effect for Use of advertising blocking software on the number of fixations on the banner ads (F (1, 58) = 1.55, p = .218). No
relationship was found between the use of adblocking software and the number of fixations on the banner ad.
Table 4. The means and standard deviations for the number of fixations on the banner ad, for users and non-users of advertisement blocking software
Advertisement blocking software n M SD
Users 42 1.18 1.43
Non-users 18 0.71 1.03
3. Proportion of duration of fixation
Additionally, attention was measured in terms of the proportional duration of
fixations on the banner ads. This part of the first research question concerned the relationship between the proportional duration of fixation on banner ads and the language of the banner ad slogan. A repeated measures analysis for proportion of duration of fixation on the banner ads with as factor Language of the advertisement slogan showed no significant main effect of language (F (1, 59) = 0.60, p = .440). No relationship was found between the duration of fixation and the language of the banner ad slogan.
Table 5. The means and standard deviations for percentage of fixations for all participants (N = 60)
Language advertising slogan
M (in %) SD
Webpage content English 93.35 0.08
Dutch 93.35 0.10 Total 93.35 0.06 Banner advertisement English 0.85 0.01 Dutch 1.30 0.04 Total 1.08 0.06
22 In general, the proportion of fixations on the banner advertisement was extremely low when compared to the proportion of fixations spent on the main content of the webpage (see Table 5).
Again, the effect of the use of advertisement blocking software on the (proportional) time spent looking at the banner ads was also investigated. A one-way ANOVA with as factor Use of advertisement blocking software (using/not-using) showed no significant main effect of the use of advertisement blocking software (F (1, 58) = 1.14, p = .291). Neither the use of advertisement blocking software nor the language of the banner ad had a significant influence on the time spent looking at the banner ad.
4. Return of views on the banner advertisements
Finally, attention was measured in terms of the return of views on the banner ads. Here, the relationship between the return of views on the banner ads and the language of the banner slogan was investigated. A repeated measures analyses for return of views on the banner ads with as factor Language of the advertisement slogan showed no significant main effect (F (1, 59) = 0.15, p = .699). No relationship was found between the return of views to the banner ads and the language of the banner ad slogan.
Table 6. The means and standard deviations for the return of views on the banner ad, specified per language of the banner ad slogan (N = 60)
Language of advertising slogan M SD
Dutch 0.46 0.68
English 0.42 0.57
Attitudes
1. Attitude toward product
The second research question concerned attitude toward several factors. First of all, the attitude toward product was investigated. The variables from the different conditions and the different banner ads were combined after the reliability was checked with the use of repeated measures analyses. Two significant results were found that needed careful consideration.
Firstly, a repeated measures analysis showed that, with regard to attitude toward product, the advertisement of the MP3-player with an English slogan differed significantly
23 from the advertisements with the smartphone and camera which had both English slogans (F (2, 56) = 6.61, p = .003). The attitude toward the MP3-player was significantly less positive (M = 3.03, SD = 1.09) than the attitude toward the smartphone (M = 3.72, SD = 1.29) and the camera (M = 3.73, SD = 1.22). Therefore, it was concluded that the advertisement promoting an MP3-player with an English slogan was evaluated differently than the other two
advertisements and that the three variables could not be combined. The advertisement promoting an MP3-player with an English slogan was therefore excluded from the analyses conducted with regard to attitude towards product.
Moreover, a repeated measures analysis showed that the attitude toward the advertisement promoting a tablet with a Dutch slogan differed significantly from the
advertisement promoting the television with a Dutch slogan (F (2, 56) = 3.96, p = .025). The attitude toward the tablet was significantly more negative (M = 3.40, SD = 3.95) than the attitude toward the television (M = 3.95, SD = 1.41). For this reason, the advertisement
promoting a tablet with a Dutch slogan was excluded from the analyses conducted with regard to attitude to product.
With the final data set, a repeated measures analysis for attitude toward product was conducted. The analysis with as factor Language of the advertisement slogan showed no significant main effect of attitude toward product (F (1, 59) = 0.01, p = .93). No relationship was found between the language of the banner ad slogan and the attitude toward the advertised product.
Table 7. The means and standard deviations for the attitude toward the advertised product, specified per language of the banner ad slogan (N = 60)
Language of advertising slogan M SD
Dutch 3.68 1.10
English 3.67 1.12
2. Purchase intention
Secondly, the second research question concerned the purchase intention. A repeated measures analysis for purchase intention with as factor Language of the advertisement slogan showed no significant main effect (F (1, 59) = 0.95, p = .334). No relationship was found between the language of the banner ad slogan and the purchase intention.
24 Table 8. The means and standard deviations for purchase intention, specified per language of
the banner ad slogan (N = 60)
Language of advertising slogan M SD
Dutch 2.61 1.17
English 2.47 1.16
3. Attitude toward brand
The second research question also concerned the attitude toward the brand of the advertised product. A repeated measures analysis for attitude toward brand with as factor Language of the advertisement slogan showed no significant main effect (F (1, 59) = 0.12, p = .774). No relationship was found between the language of the banner ad slogan and the attitude toward the advertised brand.
Table 9. The means and standard deviations for attitude toward the advertised brand, specified per language of the banner ad slogan (N = 60)
Language of advertising slogan M SD
Dutch 3.22 0.98
English 3.25 0.94
4. Attitude toward advertisement
Finally, the second research question concerned the attitude toward the
advertisements. In order to check whether the advertisements were perceived differently and whether these differences in perception caused different attitudes, a repeated measure analysis was conducted with as factor Advertisement. The analysis showed no significant main effect (F (3.27, 193.08) = 1.21, p = .309). Participants showed similar attitudes toward all the advertisements.
Table 10. The means and standard deviations for attitude toward advertisement, specified per banner ad (N = 60)
Banner ad M SD
1 3.29 1.32
2 3.06 1.21
25
4 2.96 1.11
5 3.09 1.04
6 3.15 1.20
Subsequently, the relationship between the attitude toward the advertisement and the language of the banner ad slogan was investigated. A repeated measures analysis for attitude toward advertisement with as factor Language of the advertisement slogan showed no significant main effect either (F (1, 59) = 1.29, p = .261). No relationship was found between the attitude toward the advertisement and the language of the banner ad slogan.
Table 11. The means and standard deviations for attitude toward the advertisement, specified per language of the banner ad slogan (N = 60)
Language of advertising slogan M SD
Dutch 3.15 0.99
English 3.03 0.99
As stated in the introduction of this study, the use of advertisement blocking software is said to indicate a low appreciation for online advertisements. This was investigated with a simple regression analysis. The analysis showed that the variable Use of advertising blocking software explained 6.3% of the variance for Attitude toward advertisement (F (1, 58) = 4.95, p = .030). The use of advertising blocking software was shown to be a predictor of attitude toward advertisement (β = 0.28, p = .030). This means that when participants used
advertisement blocking software, their attitude toward the advertisement went down with 28% independently of the language of the advertising slogan. It was found that the use of ad
blocking software indicates a lower appreciation for online advertisements.
Table 12. Regression analysis for use of advertisement blocking software as predictor for attitude toward advertisement (N = 60)
Variable B SE B β Use of advertisement blocking software 0.55 0.25 0.28* R² 0.063 F 4.95* *p < 0.05
26 Recall
The third research question concerned the recall of the product type, the brand name of the advertised product and the slogan used in the banner ad. The recall was measured with a free recall memory test and the responses of the participants were coded as follows:
(1) No recall.
The participant did not remember anything about the component of the advertisement. (2) Incorrect recall.
The response of the participant was incorrect. (3) Correct recall for a different advertisement.
The response of the participant was correct but did not belong to the respective webpage.
(4) Recall of the job advertisement.
The participant confused the product advertisement with the job advertisement and responded corresponding to the job advertisement.
(5) Correct recall.
The participant remembered the component of the advertisement correctly.
Table 13. Frequencies of responses given by the participants to the recall questions (N = 180)
English Dutch
Product Brand Slogan Product Brand Slogan
% % % % % % No recall 161 89.44 169 93.89 180 100 154 85.56 171 95.00 179 99.44 Incorrect recall 3 1.67 3 1.67 0 0 5 2.78 0 0 1 0.56 Recall different advertisement 5 2.78 1 0.56 0 0 1 0.56 1 0.56 0 0
Recall of the job advertisement
10 5.56 7 3.89 0 0 20 11.11 8 4.44 0 0
Correct recall 1 0.56 0 0 0 0 0 0 0 0 0 0
The score on correct recall was extremely low (see Table 13). Participants were frequently confused about this section of the questionnaire when they did not recall seeing the product advertisement at all. Therefore, the product advertisement was sometimes mistaken for the job advertisement, which led to responses concerning the job advertisements. In order
27 to be able study the data correctly and to compare the recall per language of the slogan, “No recall” and “Recall of the job advertisement” would need to be excluded. Unfortunately, this would lead to a sample size that was too small for any further analyses. Subsequently, no conclusions could be drawn.
Appreciation
The fourth and final research question concerned the appreciation for advertisements in Dutch and advertisement that contain English text elements in general. A paired-samples t-test showed a significant difference between appreciation for advertisements in Dutch and appreciation for advertisements that contain English (t (59) = 2.64, p = .011). Appreciation for advertisements that contain English (M = 4.11, SD = 1.01) was shown to be higher than appreciation for advertisements exclusively in Dutch (M = 3.77, SD = 1.04). It was found that advertisements that contain English are more appreciated than advertisements that are
exclusively in Dutch.
Table 14. The means and standard deviations for appreciation for advertisements exclusively in Dutch and appreciation for advertisements that contain English (N = 60)
M SD
Advertisements in Dutch 3.77 1.04
28
Conclusion/Discussion
The goal of this research was to gain insight into the effect of the use of English in Dutch online advertising. Therefore, an experimental study was set up to investigate the effect of English on the effectiveness of drawing attention of Internet users, on their attitudes and on their recall. Additionally, the general appreciation of Dutch online advertisements with and without the use of English was studied. The research was operationalized with the use of an eye tracking study and a questionnaire.
Attention
The first research question addressed the relationship between the use of English in Dutch online advertising and the effectiveness of attracting attention. In order to investigate this relation, attention was measured through four different variables: first fixation, number of fixations, duration of fixations and returns of views on the banner ads. For none of these four measures a significant difference was found between advertisements with English slogans and advertisements with Dutch slogans. Advertisements with English slogans did not cause
participants to focus on the banner ad firstly more often than advertisements with Dutch slogans did. Also, advertisements with English slogans did not attract higher numbers of fixations than advertisements with Dutch slogans, nor did English slogans lead to a longer viewing time of the advertisement than the advertisement with Dutch slogans. Finally, the use of English slogans did not lead to a higher number of returns of views on the advertisements than the use of Dutch slogans did. This leads to the general conclusion that in this study English slogans were not more effective in terms of drawing attention than Dutch slogans, but it can also be concluded that English slogans were not less effective either.
In general, the outcomes of the eye tracking section of this study showed that the banner ads were barely paid attention to. On average, the participants were looking at the advertisements for only 1.08% of the total time spent on looking at the webpages. This might be explained by different factors. Firstly, the banner ad obviously consisted of only a small part of the total webpage that the participants were exposed to. But most importantly, participants were asked to look for specific information that they could only find in the contents of the job advertisement. Moreover, the participants were likely to know exactly where this information could be found after having seen multiple webpages. Thus, the need to
29 look at other elements of the website only decreased in time. These factors might have led to a highly natural situation in which the participants, similarly to any Internet user, did not pay attention to anything else on the webpage than the parts of the webpage that were relevant for their goal. This finding sheds light on the actual effectiveness of online banner advertisements in general, which seems to be rather low.
The non-significant outcomes of this part of the current study with regard to the effect of language do not support those of the study of Toebes (2013). The latter indicated that English attracts more attention in the case of number of fixations and duration of fixations than Dutch. These differences can be partly explained by the mere fact that Toebes’ research concerned print advertisements and that he did not place his advertisements within a specific context, the way that the advertisements in the present study were placed within an actual webpage. The webpage additionally contained many more components for the participant to look at, which might have decreased the effect of the language of the banner ad slogan. However, it can also be debated whether the experiment in the present study did not offer a much more valid set-up. Advertisements, whether they are print or online, are almost always part of a broader context in which they assume a peripheral role and are not as prominent as in experiments in which they appear isolated from their contexts. The goal-directed task that participants received reduced the attention that participants paid to the advertisements (Danaher & Mullarkey, 2003). Therefore, the set-up in this study might resemble actual conditions more closely than Toebes (2013), since Internet users usually visit webpages with a goal: finding information. This online behavior is very similar to the behavior that was produced in the experiment of the present study.
Attitude
The second research question concerned the relationship between the use of English in Dutch online advertising and the attitudes toward the advertised product, the advertised brand and the advertisement itself. The language in the banner ads did not make a difference for any of the attitudes. Therefore, it can be concluded that the use of English slogans in the Dutch advertisements did not have any effect on the perception of the advertised product, the advertised brand or the advertisement.
The study of Toebes (2013) resulted in similar findings, namely that English and Dutch did not differ with regard to attitudes. Here, Toebes’ study and the present study support each other. However, the findings differ greatly with the suggestions that were made in the literature about the use of foreign language in advertising. According to De Raaij
30 (1997), the use of English in Dutch advertisements would make advertising texts appear more intelligent and interesting. This could be expected to lead to a more positive attitude toward advertisements with English. Additionally, Takashi (1990) suggested that English would lead to more positive attitudes.
A general explanation for the lacking effects of English in the present study could be that Internet users are used to encountering English on the Web, as the Internet is such an international medium. Moreover, the strategy of using English in advertising has been used so frequently by advertisers (Gerritsen et al., 2007) that it has been hypothesized that the use of it does not stand out anymore (Planken et al., 2010). The eye tracking study of Toebes (2013) also supports this suggestion. In other words, the combination of the international features of the Internet and the current prevalence of English in multiple genres of advertising, seem to make English barely distinctive for the non-English audience of online advertisers. This finding hints that the influence of English as a dominant majority language on the Internet is much larger than the influence of local oriented advertising. Targeting online audiences with personalized messages in their L1 with the use of Online Behavioral Advertising may have only little effect on the actual perception of the advertised message.
Recall
The third research question addressed the relationship between the use of English in Dutch online advertising and the recall of the advertised product, the advertised brand and the
advertising slogan. This relationship could not be investigated properly as, generally, the measured recall was too low. The low recall was caused by the lack of attention that participants had paid to the banner ads. In some cases this caused misunderstanding and confusion about the recall questions. During the filling out of the questionnaire, some participants even consulted the researchers in order to check whether something had gone wrong during the eye track experiment.
The low recall was not surprising as it was already concluded that the participants rarely looked at the banner ads during the eye tracking experiment. With the low values for
proportional duration of fixation on the banner ads and the number of fixations on the banner ads, the recall could have been expected to be this low. Moreover, Danaher and Mullarkey (2003) found that goal-directed behavior of web users leads to a lower recall of banner ads than exploratory behavior. Unfortunately, this outcome did not yield any insights into the relationship between language and recall.
31 Appreciation
The fourth research question concerned the preference of participants for advertisements that are exclusively Dutch and advertisements that also contain English. Participants reported a significant higher preference for advertisements containing English than for advertisements that are entirely in Dutch. Seemingly, participants perceive their own interest for online advertisements with English higher than their interest for online advertisements that are exclusively in Dutch, which is in great contrast with the findings on the actual attitudes in relation to the different languages of the banner ad slogans. The difference between the actual attitudes and the perceived preference of the participants is interesting. Participants indicate a higher appreciation for advertisements that contain English, but when attitudes are being researched, no such difference is found.
Additional findings
Finally, the influence of the use of advertising blocking software was investigated since its popularity has been increasing rapidly. It was hypothesized that the use could lead to a change of looking behavior, as Internet users are less used to seeing online advertisements and could be more prone to spot them. Also, the use of ad blocking software could indicate a more negative attitude toward advertisements. In this study, no significant differences were found in looking behavior between participants that reported to use adblocking on their computer or laptop and participants that did not. This finding implies that less exposure to online
advertising did not lead to a different response when the participants in this research encountered an advertisement banner.
However, the results of this study do show that participants that have installed adblocking software on their computer or laptop have a significantly more negative attitude toward advertisements than the participants that have not installed the software. This finding supports the suggestion by PageFair and Adobe (2014) that users of adblocking software experience great annoyance with advertisements. Internet users that experience high amounts of annoyance with online advertisements might be more likely to take action and install adblocking software.
Limitations, implications and further research
When evaluating this study, one may conclude that there are some limitations that affect the usability of the findings. Firstly, the experimental setting that included the attached head band may have caused the participants to feel slightly unfamiliar and uncomfortable.
32 Moreover, a habituation effect may have occurred because the participants had to look at six webpages that were all similar in layout. The ease of fulfilling the task might have increased after each webpage while the curiosity and attention for the other components of the webpage might have decreased.
However, a very strong aspect of this study was the validity of the experimental setting that was used. Due to the goal-directed task that participants received, the results are likely to match the actual looking behavior that Internet users show closely. Therefore, the research has led to several implications that could be used in practically relevant situations and that can be the basis for further research. Additionally, the low effectiveness of banner ads that was found in this study raises questions with regard to the validity of advertising research that lacked the use of goal-directed tasks or close to real-life experimental settings.
Overall, this study implies that the use of English does not attract more attention in Dutch online advertising and that it does not cause a more positive attitude toward the product, the brand or the advertisement. Even though the target audience thinks English makes Dutch online advertisements more appealing, the actual effect is lacking. Nevertheless, the outcomes also imply that English is not less successful than Dutch either, with regard to drawing attention and to the attitudes that the advertisements evoke. Moreover, this study indicates that Internet users that browse the Internet for information rarely pay attention to banner advertisements. This lack of attention does not differ for Internet users that have installed adblocking software.
This study was just a small step into a field where more research could provide many more useful and applicable insights. Future research could improve the validity of this study by adding animations and interactive aspects to the online advertisements, as this could make the design of the experiment even more realistic and vivid. Additionally, an interesting topic for more research would be the discrepancy that was found in this study between the
perceived preference for the use of English in online advertisements and the lack of effect of English on the attitudes evoked by the advertisement. More research into this discrepancy would be valuable, as it seems that the participants in this research did somehow attribute positive qualities to English, as the literature suggests (e.g. Hornikx & Starren, 2006; Piller, 2001; Van Meurs, 2010; De Raaij 1997). Interviews could be used to investigate the
background of this positive attitude toward advertisements that contain English. It could also be investigated whether this discrepancy remains over time, as the effects of English seem to decrease over time due to the factors that were mentioned earlier.
